A story of the legendary 25W "Otala" amplifier as seen by Terje Sandstrøm, one of the original designers.
Also see "The People Involved", for other designers and people who contributed.
Introduction -The birth of the Otala amplifier
On an AES conference in 1973, Dr. Matti Otala presented a paper describing the design of a TIM-free audio amplifier. Present at this conference was Svein Erik Børja, a Norwegian record and broadcasting producer, and a great audio enthusiast. Svein Erik Børja also was one of this times greatest Golden Ears. He was able to hear even the slightest imperfection in an audio component. Having been dissatisfied with the sound of that days transistorised audio amplifiers, he here saw an opportunity for getting an audio amplifier of a new generation. The talk of Dr. Otala about TIM also was an explanation to Svein Erik for the imperfections he himself had noted in audio amplifiers.
He brought the paper to a friend of him who was running his own audio company in Norway, Per Abrahamsen of Electrocompaniet. Per decided to make a try at the amplifier to see what it might bring. They got the help of Nils Jørgen Kjærnet at Nera in Oslo, who made circuit boards design and also, to my knowledge, helped in the mechanical design of the amplifier.
They first made just a couple of amplifiers, but the sound was so good, fulfilling all their expectations, they decided to manufacture a series of these amplifiers.
I started showing up at Electrocompaniet in the autumn 74, and started to work for Per in the spring of 75. One of my first jobs was to assemble the first production series of the Otala amplifiers. It was a series of 10 amplifiers, based on the same PCB's as the original two prototypes. During the summer I started to look at the design, having been an electronics hobbyist since early teenage. At that time neither Per nor myself knew too much about high-end audio design. It may well have been the factor that brought forth the success of the amplifier, as it evolved during the next 5 years. We didn't know how it should be done, so we did it out from the framework given by the Otala amplifier. We may well have been the first audio amplifier designers in the new school - the TIM free designers, so to speak.
The first Otala amplifier in "production"
The first amplifier series was using the same printed circuit boards as was used on the first two prototypes. The PCB was a two-layer design, we called it a T-board, due its shape. The power transistors were mounted onto the cooling fin from both sides of the stem of the 'T'. One layer of the board was a ground plane (Nera being a RF-company, it was not so strange.) The ground plane caused us a lot of mounting problems, because the component legs often got stripped when they were squeezed through the holes in the PCB, causing small metal pieces from the legs to curl up and making sweet, nearly invisible, shorts to the ground plane. We did see smoke....
In the autumn 75 we started to get some attention, and a visit to Matti Otala at the VTT (Valtionen Teknilinnen Tutmiskeskus - Technical Research Center of Finland ) where he worked as a professor was needed.
There were several reasons:
We had named the amplifier "Otala - Lohstroh" after its inventors, and Otala wanted to see the amplifier before he and Lohstroh possibly could allow us to use his name.
At the same time, we started to experience problems with the amplifier - it didn't quite have the specifications stated in the AES paper. We needed his help.
As it turned out, the amplifier was renamed to "The 2 channel audio power amplifier", which was the name used for the rest of the amplifiers lifetime. The relationship with Otala was established, and he visited the company on several occasions. There is still a good contact between Otala and Electrocompaniet.
Period of confusion, start of revision
It was early 1976. We started to invest in measuring equipment. The company turned slowly into a audio amplifier company. The company lost several of its old customers, the loudspeakers that the company were selling was more neglected. We worked from early morning into the late nights. That should be the standard for the coming 4 years. An amplifier brought to the US by Svein Erik blew up, at the first listening test - embarrassing. We used the night to invent a new type of short circuit protection. We didn't want anything that could affect the sound, so we simply used a high impedance circuit to sense the current through the output transistors, and then used a relay to switch of the voltage to the output transistors. It worked!
We now understood that we had to start changing the amplifier. Svein Erik's golden ears got even more important than they had ever been before. We were on uncharted territory, the old measurement methods could not be trusted. We had to rely on our own methods, our own interpretations of the measurement results. Now started an active period with theory, practice, reading articles (Svein Erik provided us with articles 'en masse', on all aspects of audio design, and electronics design in general.)
We found that the compensation scheme used in the amplifier did not work correctly. As we measured the bandwidth to be far less than what was specified, we needed to redesign these filters. The amplifier used shunt-feedback, one of its really strong points, with an input lag compensation. At each amplifying stage, there was three, a lead-lag compensation was introduced. These did not match the actual pole's we measured. One of the reasons for this, we assumed was that the original design was using "fresh" newly developed Phillips transistors. As both Otala and Lohstroh were working at the Eindhoven Phillips lab's at the time of the design we assumed they could have access to better transistors. (I don't remember if Otala ever confirmed this to be case.)
In any case, a period started where we changed the compensation back and forth with no or little effect. We did achieve something, but not any significant improvement.
We had also now measured the distortion (having just bought our first distortion measurement set and spectrum analyzer). It was not nice, the distortion was far worse than specified, and far below even our expectations. Again we didn't manage to do anything significant at first.
A major change happened first when we started to change the quiescent current of the transistors. The design was in fact done as a mix between the old way and the new way. Older design books learn that you should use low current in the first stage do to noise. This was also done here, but (see "Noise optimisation") for an explanation of why this is not the case). We knew this theory to be in error, and increased the quiescent currents, thus decreasing the resistance and again increasing dramatically the bandwidth of the amplifier. The distortion also went down, not dramatically but enough to make a significant change in the sound to the better (see "Transistors, resistors, current and distortion" [coming]).
The Audio Critic
At this time we started to hear rumours about a test in the US on our amplifier. One day our mail started to overflow, and a few days later we got hold of the test ourselves. It was in a magazine named The Audio Critic and the test was fabulous. It started "Audio freaks - Eat your hearts out: This is the worlds best sounding amplifier, and One: You can't buy one in this country (The USA!), and Two: (not surprisingly - the only slightly negative in the whole test) It is too low powered to be counted. " Suddenly we had more requests for amplifiers than we could hope to handle. Sales boomed. It was mid 76.
We worked all summer, mornings - days - evenings - and into the night. Monday to Friday, Saturday and Sunday. Per's wife, Anne, was not always too happy about this. Per had two small boys, and she had to take care of them a lot by herself. Anyway - her "sacrifice" made the makings of these amplifiers possible, and a good thing about it all is that both Per's and Anne's sons are now working at Electrocompaniet ! So at last it turned out to be an investment into their future as well !
Dr. Otala's theorems on TIM very often came out as an attack on high-feedback amplifiers. Although it is correct that high-feedback amplifiers are more prone to TIM than low-feedback amplifier, there is no magic here. The main goal of any amplifier is to reproduce the incoming music as perfectly as possible, not adding nor subtracting anything. TIM is just one type of distortion, it will not help to exchange one type with another. It is true that some types of distortion are harsher to the ear than others, but there still is a balance to be done. If 0.1% of TIM equals 1% of THD in audibility, then an amplifier with 2% THD and 0.07% TIM will sound worse than an amplifier with 1% THD and 0.1% TIM.
So, we realised than the balance of distortions were the essential factor to consider. Not only do you have to balance THD against TIM, but also low frequency distortion against high frequency distortion, frequency and phase response against non-linear distortion in general, and so on.
This insight triggered the Great Change:
One night (it always happened at night times!) we increased the feedback 10dB, to a total amount of 30dB feedback. The sound improvement was staggering!!! And, contrary to common belief in our own community !
After this we only adjusted the amplifier slightly. It had found the form it should have for its remaining life at Electrocompaniet. And, the amplifier was a success!
On Marketing
We did brochures ! The first one is seen here. Second one is here.. The rest of the marketing was done by ear-to-ear rumours. The fact that the amplifier and the company was sort of hard to get stimulated the market. Here was a mysterious company with a killer amplifier. Audio people love such stuff !
Technical things
See the schematics. Also, see some of the calculations that were done.